Old Sinitic "rice", with an added note on "leopard"

July 9, 2023 @ 6:38 am · Filed by under Historical linguistics, Language and biology

« previous post | next post »

We've had extensive discussions about the Old Sinitic reconstruction of the Sinitic word for "wheat".  Although we've been circling around it for quite some time now, we haven't yet nailed it down securely, but we're close.  While we're still occupied with "wheat", Martin Schwartz sends in this terse, seemingly cryptic, but extremely interesting information about words for rice:

Sorry I can't help by citing the reconstruction I saw in Boodberg

which looked like it was compatible with PIIr. *wrinźh.
 
(6/13/23)

Before digging into the implications of PIIr. *wrinźh for our ongoing quest to find archeolinguistic links between eastern and western Eurasia, I'd like to say a few words about Peter Alexis Boodberg (1903-1972), whose hallowed name has come up several times on Language Log (see here and here [vigorous discussion in the comments]).

Sinology has more than its share of eccentric geniuses (see E. Bruce Brooks, "Sinological Profiles", for a couple dozen), and Boodberg certainly was one of them.  A professor at Berkeley, he had a colorful, eventful past stretching back to 13th c. Baltic German nobility (and before that to the city of Mainz in 1003) and birth in Vladivostok to a father who was a baron and the commanding general of the Russian forces there.  Boodberg had a fecund mind (and fertile imagination!).  I find some of his insights to possess great explanatory power, while others are quirky and quixotic:  his ideas about dimidiation, where a phonologically complex syllable could be split into two Sinitic morphosyllables, to me, at least, belong in the former category, while his notion that Sinitic etymons could be systematically compared to Latinate roots falls in the latter category.

If you are interested in learning more about Boodberg's scholarship, try to find his self-printed Cedules from a Berkeley Workshop in Asiatic Philology.  For easier access to the products of Boodberg's scintillating mind, turn to the following two works:

Cohen, Alvin P., ed. (1979). Selected Works of Peter A. Boodberg. Berkeley: University of California Press.

Honey, David B. (2001). Incense at the Altar: Pioneering Sinologists and the Development of Classical Chinese Philology. American Oriental Series. Vol. 86. New Haven, Connecticut: American Oriental Society.

Boodberg's own single most famous work is arguably this 1937 publication:

"Some Proleptical Remarks on the Evolution of Archaic Chinese". Harvard Journal of Asiatic Studies. 2.3-4: 329–72. doi:10.2307/2717943. JSTOR 2717943.

Martin Schwartz received his PhD in Iranian studies from Berkeley in 1967 while Boodberg was still teaching and had numerous conversations with him, so is familiar with the latter's Sinological researches and has a good sense of their sources and value.  Given that the seemingly cryptic comment above comes from Martin, who is generally skeptical of such things (long distance connections), as we know from his many animadversions on Language Log, we needs must take it seriously.

Apud Martin, we don't know exactly what reconstruction Boodberg envisaged, but we do know for sure that it evoked a resonance with PIIr. *wrinźh in Martin's mind.

Martin says that he knew Boodberg and much enjoyed talking with him.

When I was a grad student at UCB, I noticed he would be speaking at the AOS (?) on Chin. 'being', a topic intellectually sexy for a young man. I took the bus down to the venue, arriving late; outside the bldg
was a gentleman whom I asked if Prof Boodberg was yet to speak; he answered, "I am Boodberg" and offered to drive me back to Bkly.  It was a great ride. I would speak to him thereafter whenever I'd see him in the coffee room we had in Dwinelle Hall.
 
Perhaps he wasn't always right in his speculations, but a wrong from Boodberg was more interesting than a right from less brilliant scholars.  I think you'll agree.  I was friendly with his student Bill Boltz, too.
 
I looked again at the comments to LL "Of shumai and Old Sinitic reconstructions"; I now see you cite a "Berkeley colleague" re a Boodberg reconstruction in his Selected Works of 'rice', which colleague thought was compatible with IIr. *wrinżh- . I think the "colleague" was probably
 
Yours Truly,
Martin
 
I asked Laurent Sagart, who has worked extensively on Old Sinitic terms for grains, especially possible connections with non-Sinitic languages.  He replied:
 
My current thinking on Sino-Tibetan cereals is summarized in the supplementary to the article "Dated language phylogenies shed light on the ancestry of Sino-Tibetan" published in the PNAS (open access, then), section 5. Here is the link to the supplementary:
file:///Users/laurent/Downloads/pnas.1817972116.sapp-1.pdf
I presented my ideas on cereals at the broader east Asian level in a book chapter in a volume on the archaeology of SEA, published 2022: https://www.academia.edu/45208849/LANGUAGE_FAMILIES_OF_SOUTHEAST_ASIA
 
You asked about the origins of "the word" for rice in OC. I suppose you mean dào 稻, the modern name of the rice plant, but originally referring to rice grains, with signific mǐ 米 where the modern character has hé 禾. Baxter-Sagart 2014 reconstruct *[l]ˤuʔ. I think this word is part of a root also including yǎo 舀 'to scoop out', and that it originally referred to rice grain in storage, still with the husks on, and ready to be poured into the mortar to be pound in preparation for cooking. There are probable TB cognates:
    • WT blug ‘fill a bowl with rice’ (< m-lug) ; ldug ma 'a scoop', ldug pa 'to pour' (ldug < 'lug)
    • Proto-Tani (Sun) lɯk 'to pour'
    • Chepang (Caughley) hluk- 'pour over, on (in quantity), sluice with water'
I think the PST did not cultivate rice, so there is no PST word for 'rice'. The other modern Chinese words for rice in its various forms: mǐ 米 lì 糲 fàn 飯, etc. are not rice-specific. They apply as well to the millets and are part of the general ST vocabulary of cereal agriculture; the first two have Austronesian cognates.
 
George van Driem called my attention to these two recent publications by him:
 
George van Driem. 2017. ‘The domestications and the domesticators of Asian rice’, pp. 183-214 in Martine Robbeets and Alexander Savelyev, eds., Language Dispersal Beyond Farming. Amsterdam: John Benjamins.

George van Driem. 2012. ‘The ethnolinguistic identity of the domesticators of Asian rice’, Comptes Rendus Palévol, 11 (2): 117-132.
 
Axel Schuessler, ABC Etymological Dictionary of Old Chinese (University of Hawai'i Press, 2007), has several entries for different types / categories of rice — e.g., bài 粺, dào 稻, mǐ 米, etc. — some of which have areal non-Sinitic comparanda, but none of which sound remotely like IIr. *wrinżh.

Coming at "rice" from the west and the south, we have:

mid-13c., ris, "edible seeds or grains of the rice plant, one of the world's major food grains," from Old French ris, from Italian riso, from Latin oriza, from Greek oryza "rice," via an Indo-Iranian language (compare Pashto vriže, Old Persian brizi), ultimately from Sanskrit vrihi-s "rice."

The Greek word, directly or in indirectly, is the source of the European words for the grain (Welsh reis, German reis, Lithuanian ryžiai, Serbo-Croatian riza, Polish ryż, etc.). Evidence of semi-cultivated rice in Thailand dates to 5,500 years ago; introduced to the Mediterranean by the Arabs, it was introduced 1647 in the Carolinas.

etymonline

From Middle English rys, from Old French ris, from Old Italian riso, risi, from Byzantine Greek ὄρυζα (óruza), from an Eastern Iranian language related to Middle Persian blnj (*brinǰ). Theorized to come to Iranian languages from Sanskrit व्रीहि (vrīhi).

Prior to Sanskrit, it is speculated to be either a borrowing from a Dravidian language (compare Proto-Dravidian *wariñci (“rice”)), thence from Austroasiatic languages such as Proto-Mon-Khmer *sruʔ (“paddy rice”).

Wiktionary

Now that matches with Martin's IIr. *wrinżh.

As for the history of rice cultivation,

The current scientific consensus, based on archaeological and linguistic evidence, is that Oryza sativa rice was first domesticated in the Yangtze River basin in China 13,500 to 8,200 years ago. Cultivation, migration and trade spread rice around the world – first to much of east Asia, and then further abroad, and eventually to the Americas as part of the Columbian exchange. The now less common Oryza glaberrima rice was independently domesticated in Africa around 3,000 years ago. Other wild rice species have also been cultivated in different geographies, such as in the Americas.

Wikipedia

None of the above assembled evidence bodes well for an external (particularly IE) connection of Sinitic words for "rice" at the earliest stages.

Last resort

I keep Boodberg's Selected Works displayed prominently on a shelf in my basement dungeon study.  It's approximately 500 pages long and typographically not very friendly, plus it doesn't have an index.  So I started to wade / plod through it in search of a word that means "rice" and sounds like *wrinżh. 

Page after page.  Finally, on p. 372, I found a discussion of nián 年, usually rendered as "year", but which can also signify "harvest".  Boodberg seems to reconstruct (based on Karlgren) the Old Sinitic form of nián 年 as *nzien <**znien (with all letters except the last having tiny diacritical marks).  The problem with this is that the phonetics do not mesh well with *wrinżh and the semantics indicate "wheat; grain; harvest", not "rice".  (Maybe I missed some other cereal word discussed by Boodberg, but this one doesn't appear to work for *wrinżh.)

Glyph origin

In the oracle bone script and early bronze inscriptions, it was originally 秂, an ideogrammic compound (會意) and phono-semantic compound (形聲, OC *niːŋ): semantic 禾 (“wheat; grain”) + phonetic 人 (OC *njin, “person”) – a person carrying wheat on his back – harvest.

In bronze inscriptions after the Western Zhou period, a stroke was often added to 人 to give 千 (OC *sn̥ʰiːn), which still acted as a phonetic component, and this form (秊) was inherited by later scripts. The current form is inherited from the clerical script, where libian (隸變) has occurred.

Etymology

From Proto-Sino-Tibetan *s-ni(ː)ŋ ~ s-nik (“year”).

Pronunciation

(Baxter–Sagart): /*C.nˤi[ŋ]/
(Zhengzhang): /*niːŋ/

(source for the above four sections)

 

Summary lesson

Like looking for a grain of rice in a haystack.

 

Leopard

Prompted by colleagues, at the same time I asked Martin about "rice", I also inquired whether he had any observations concerning "leo_pard_".  He replied:

Am I to comment on the leo or the pard?
 
For 'lion' I had some ideas in my old article "Gathic Compositional History…", which is on my (underrepresentative) page on academia.edu; the discussion is on pp. 263-5.  PLEASE NOTE THAT MY POINT OF DEPARTURE IS WRONG; FOR sarjan- I LATER PROVED THAT THIS IS *sar-jan- "smasher of bonds', AND NOT FROM A ROOT *SARG/J- !!!  Also,  the Germanic words I treat are from PIE √*sk'elg/g', with palatal k'; I was given no chance to read the proofs of this article.  I no longer especially favor e.g. √haiz here. i leave Tocharian lions on whatever they lie on; one could look at DQ Adams 1984 for that, and re lion him more than rely on me. Nor will now enter the second-hand cage containing possible Sinitic connections, which is fraught with fright for me. Off my map; "hic sunt leones".
 
As for the pard part, I have nothing interesting to say. Surely Greek párdos etc. represent an easterly loan; Proto-Iranian "prdangā- (with syllabic r) regularly gives Sogdian pur∂ang, Pashto pRāng, and Pers. palang. OInd. prdāku- (again with syllabic r) is obviously related. Pard-on my lack of anything new to say. But I have nothing to say on the question which may most interest your readership, the possible relationship(s) between the Iranian, Old Indic, and Chinese 'lion' words. I think Boodberg as well as Henning were interested in this.
 
Classical Arabic has scores of words for 'lion'–indeed, any truly literate old-school Arab will proudly rattle off at least ten.  See David Larsen, Names of the Lion for a poem listing the various names.  None is remotely like the ones at question!
 

Selected readings

[Thanks to Jichang Lulu and Chris Button]

July 9, 2023 @ 6:38 am · Filed by under Historical linguistics, Language and biology


41 Comments

  1. David Marjanović said,

    July 9, 2023 @ 8:33 am

    Here is the link to the supplementary:
    file:///Users/laurent/Downloads/pnas.1817972116.sapp-1.pdf

    As you can see from the "file:///" part, this is a link to Prof. Sagart's own harddisk that nobody else has access to.

    The paper is indeed in open access here on the PNAS website, and the supplementary information is likewise publicly accessible here.

  2. Chris Button said,

    July 9, 2023 @ 9:01 am

    I found a discussion of nián 年

    Here I reconstruct 年 ᶮʝǝmˑ and 稔 ᶮʝǝˑmɁ

    The association is with Old Mon OM cnam “year” after Benedict (1976).

    The -m coda follows Pulleyblank's (1995). For the onset, note the Shiming sound gloss of 年 ᶮʝ- with 進 c-.

    + phonetic 人 (OC *njin, “person”) … In bronze inscriptions after the Western Zhou period, a stroke was often added to 人 to give 千 (OC *sn̥ʰiːn)

    Here I reconstruct 千 scjǝmˑ ~ cʰjǝmˑ

    I think Schuessler's tentative association with Old Mon lŋim “thousand” may be correct.

    Worth noting that Pulleyblank (1995) speculatively suggested a possible -m in 人 and an association with 男.

    In my system, that would be 人 ᶮʝǝˑm and 男 nǝmˑ.

    The phonological differences are too much for a direct etymological association, but they are perhaps not too far for a common external origin if we note Schuessler's (2007) suggestion of a connection of 男 OM ɲum “prefix to name of slave master”, kɲum “young, child”

    lì 糲 … Austronesian cognates …

    … bài 粺

    The Austronesian cognate is well known here, but we should note that it is the source of both 糲 and 粺/稗.

    We've had extensive discussions about the Old Sinitic reconstruction of the Sinitic word for "wheat".

    I still stand by what I posted here, albeit to little acclaim:

    https://languagelog.ldc.upenn.edu/nll/?p=59111

    麥 ʁək and a comparison with Proto-Indo-European rugh- (or rw̩kʱ-) "rye".

  3. Chris Button said,

    July 9, 2023 @ 10:56 am

    The Austronesian cognate is well known

    At least as far back as Maspero 1933, I believe.

  4. Doctor Science said,

    July 9, 2023 @ 12:18 pm

    Better than Wikipedia is this (open-access) article, A Journey to the West: The Ancient Dispersal of Rice Out of East Asia, though van Driem (2017) is also excellently nuanced.

    All the research emphasizes that rice was domesticated and initially cultivated by non-Sinitic speakers. I wonder if the focus on the Yangtze valley that van Driem deprecates is an effort to make rice domestication part of the Glorious History of the Han Minzu–as they've done with silk IMHO.

  5. Chris Button said,

    July 9, 2023 @ 2:46 pm

    As for the pard part, I have nothing interesting to say. Surely Greek párdos etc. represent an easterly loan; Proto-Iranian "prdangā- (with syllabic r) regularly gives Sogdian pur∂ang, Pashto pRāng, and Pers. palang. OInd. prdāku- (again with syllabic r) is obviously related.

    豹 *praqs "leopard" doesn't seem too far off.

  6. Jonathan Smith said,

    July 9, 2023 @ 5:51 pm

    Hmong-Mien rice vocabulary is rich and in Ratliff's (2010: 242) view largely indigenous; that Chinese speakers adopted some of the relevant terminology and technology from HM groups as they spread southwards makes a great deal of sense. In particular, (Ratliff's) PHM *mbləu 'rice plant' has often been compared to Chin. (MC) dawX 稻 (thought to be from an earlier lateral [cluster?] onset; cf. at Schuessler [2007: 207-208].) I don't think this and similar items really reconstruct to PMin within Chinese, but rather look a lot like expansion-era loans into "early southern Chinese" > other Chinese.

    Re: Yangtze valley (the HM homeland), not only Han chauvinists seem happy to portray this and the broader south as part of a fundamentally Chinese cultural sphere: ironically (or something?), Baxter & Sagart (2014) have come to treat vocabulary shared between Chinese and the southern languages as of Chinese origin basically across the board (see Sagart's quoted remark on dào 稻) — indeed, treat the relevant etyma as they exist in Viet., HM, etc., as literal Sinitic daughter material. Bemusing…

  7. Victor Mair said,

    July 9, 2023 @ 8:01 pm

    From Martin Schwartz:

    Thanks for taking my old remark so seriously, but I'm embarrassed and frustrated.. I don't have the Selected Writings anymore (my copy had my ref. to the word on the flyleaf, so I don't know what relevant entry I saw. Living in Mount Shasta now, I cannot check; I no longer have connections with the UCB library, and cannot as anyone to go thru all the chaff to find that putaive grain, i.e. wringe (ahem!) the bogged paddy for it. Maybe someday during a protracted visit to Berkeley I can check. Until then, it's best to forget the matter.

  8. David Marjanović said,

    July 10, 2023 @ 2:11 pm

    https://languagelog.ldc.upenn.edu/nll/?p=59111

    麥 ʁək and a comparison with Proto-Indo-European rugh- (or rw̩kʱ-) "rye".

    I left a comment there doubting the existence of such a root in PIE.

  9. Doctor Science said,

    July 10, 2023 @ 2:57 pm

    @Jonathan Smith:
    not only Han chauvinists seem happy to portray this and the broader south as part of a fundamentally Chinese cultural sphere
    Exactly! In fact I, a fumbling amateur, hadn't realized that the Hmong-Mien homeland was as far north as the Yangtze valley, I thought it was in the Nanling Mountains & further south.

    I'm currently reading "Critical Han Studies: The History, Representation, and Identity of
    China's Majority" (2012), and it almost seems to me as though the formation (solidification) of Han identity "against" northern & western outsiders is misdirection, so we don't look at how fluid (absorptive?) it is to the south & east.

    Even as a very naive outsider, the standard images of "the spread of civilization in China"– where it starts in the mid-Yellow River Valley, then fills the Central Plain, the whole Yangtze Valley, then through the mountains and down to the Pearl River Valley–have always struck me as preposterous. How could the Yangtze & Pearl valleys be "there for the taking"?

  10. Chris Button said,

    July 10, 2023 @ 10:52 pm

    A little more on Austronesian *bəʀas and Maspero's (1933) proposed connection with Old Chinese …

    Pulleyblank "Old Chinese *-s" (1973) notes how OC -s occasionally hardened dialectally to Middle Chinese -t (the two readings of 四 as EMC siʰ ~ sit being the classic example). Separately Schuessler (2007, 2009) treats -ts (i.e. -t with an -s suffix) in certain cases as plain -s.

    I think Schuessler is spot on in suggesting that the difference he is noting results from root final -s versus suffixal -s, which is something attested elsewhere in Tibeto-Burman languages but generally not reconstructed for OC. (Incidentally, I should note that I disagree with Schuessler's related suggestion for an OC -h ~ -s variation, which I think stems from his failure to recognize a -ɣ coda as a counterpart to -k)

    So armed with Schuessler's root final -s and Pulleyblanks dialectally hardened -s (which I would also venture is limited to root final -s), we can look at 糲 with its three EMC readings liajʰ ~ lajʰ ~ lat. The variation automatically suggest external influence. We can call on Schuessler for OC *raˑs ~ *ras (rather than *raˑts ~ *rats) for the first two readings, and Pulleyblank for the hardening of the coda in OC *ras to -t for the third reading.

    We can then add in 稗 EMC baɨjʰ for the bilabial onset, which we can reconstruct as OC *bras (rather than *brats, or even *brajs if misled by the written phonetic component) to parallel Austronesian *bəʀas.

    @ David Marjanović

    I left a comment there doubting the existence of such a root in PIE.

    Thanks so much for posting. It seems like there is some Iranian and Welsh support to accompany the evidence from the Northwest. That's not to deny that it could still be a Wanderwort though.

  11. David Marjanović said,

    July 11, 2023 @ 12:55 pm

    "Rice" is in fact /bras/ in Balti, the westernmost kind of Tibetan…

  12. Doctor Science said,

    July 11, 2023 @ 1:52 pm

    While we're at it — are the many technical terms associated with silk production definitely of Sinitic origin and not borrowings? Silk is mythologically attributed to the Yellow Emperor (Yellow Empress?) and the Central Plain, but AFAIK the leading silk-producing regions have been the warmer, wetter rice-producing areas at least since the Han dynasty. So I wonder if silk, like rice, is a bio-technology that was first developed by non-Sinitic peoples.

  13. Victor Mair said,

    July 11, 2023 @ 2:34 pm

    Thanks to Michael Carr, her is The Selected Works of Peter A. Boodberg:

    https://www.dropbox.com/scl/fi/zmo0ne2dchzbgfywaszhl/Cohen-Selected-Works-Peter-Boodberg.pdf?rlkey=jona66j2ly8qpjazkea636buf&dl=0

    He says:

    I haven’t used Dropbox in a while, but think you can click this link, close the login box, and click Download at the upper left.

  14. Chris Button said,

    July 12, 2023 @ 6:34 am

    @ David Marjanović

    In addition to Tibetan, it’s also in Kuki-Chin as *rɐs “fruit” where the evolution of the coda to a glottal stop in some languages, rather than a tone contour indistinguishable from suffixal -s, show clear evidence of the -s as an original coda.

  15. Chris Button said,

    July 12, 2023 @ 8:59 am

    Actually, I should say " rather than a reflex conditioned by a suffixal -s" instead of "rather than a tone contour indistinguishable from suffixal -s" since tone contours were not always the reflexes of suffixal -s (the evolution of open syllable with suffixal -s being a case in point).

  16. Jonathan Smith said,

    July 12, 2023 @ 10:38 am

    @Doctor Science
    Yeah — spread of Sinitic speakers to the south/east has been spread out over ~3000 years and across countless micro-vectors, I guess yielding kind of a cultural (asnd linguistic) continuum within which it's often tricky to attribute particular features to this or that early contributor. Agree re: silk. Also paper?

    Re: lì 糲 FWIW, no idea how one could conclude that this "word" is part of modern Chinese… who ever says/said it or knows what it "means"? And upon inspection, even early written works often have things like li4mi3 糲米 li4fan4 糲飯, suggesting this item meant 'coarse'.

  17. David Marjanović said,

    July 12, 2023 @ 11:05 am

    in Kuki-Chin as *rɐs “fruit”

    Oh, interesting.

  18. Jonathan Smith said,

    July 12, 2023 @ 12:41 pm

    One of the important "voiced aspirated" items of Norman's Proto-Min is called 'barnyard grass~millet / cockspur grass / tares' and (usually) represented "稗" in the literature. E.g. Taiwanese phue7, Proto-Min "*bh-". As is again and again the case, it's quite unclear that this can be taken to proto-Chinese proper; it could be a regional item that entered mainstream Chinese at a relatively late date. More work on so-called "*Dh-" is called for…

  19. Doctor Science said,

    July 12, 2023 @ 4:19 pm

    @Jonathan Smith:

    I don't have the skills to compare the technical terms for silk and paper making across the relevant languages, though it might be a question to set some of your students looking into.

    I don't know how much of a linguistic difference it makes that textile manufacture was strongly gendered. If paper-making started as an outgrowth or sideline of textile production, it, too, may have been "women's work" for a long time.

  20. Philip Anderson said,

    July 12, 2023 @ 5:28 pm

    @Chris Button
    Welsh rhyg "rye": GPC suggests this was borrowed from Old English ‘ryge’:
    https://www.geiriadur.ac.uk/gpc/gpc.html

  21. Chris Button said,

    July 12, 2023 @ 9:04 pm

    @ Philip Anderson

    That's what David Marjanović suggested over on the thread there too:

    https://languagelog.ldc.upenn.edu/nll/?p=59111#comment-1606949

    Although it is noted by the authors to be "irregular". In any case, the Iranian evidence is still there.

  22. David Marjanović said,

    July 13, 2023 @ 10:56 am

    In any case, the Iranian evidence is still there.

    Yes, but as you quoted, it looks like a Pre-Slavic loan into Vaguely Northeast Iranian or even just a part thereof. (For pragmatic reasons that would make a lot of sense.) That may make it too recent to be passed on into OC.

  23. Chris Button said,

    July 13, 2023 @ 11:35 am

    Maybe so, and we are talking about oracle-bone time depth. But who can say for sure. It is in the vicinity, works phonologically and came from somewhere at some point.

    At least from internal evidence across the OC lexicon, I certainly challenge the notion that there was originally a bilabial m- onset. So looking for comparable forms with m- is probably not going to yield anything.

  24. Chris Button said,

    July 13, 2023 @ 11:54 am

    It would also be interesting to hear from Mallory & Adams on the possible Pamir association (where Kroonen et al suggest an alternative) since that's Tian Shan territory.

  25. Jonathan Smith said,

    July 13, 2023 @ 2:54 pm

    @Doctor Science — thanks for the interesting observation re: textile manufacture…

    Re: 'wheat', this is adduced as a straightforward Proto-Min *m- word by e.g. Norman (1973: 234). Fairly significant internal evidence… and it should go without saying that comparative procedures on mainstream Chinese would also yield *m-.

    Fuzhou Min readings of e.g. characters like "維" or "惟" are not relevant to Chris Button's hypothetical early Chinese development *ʁ- > m-. Such "words" /mi/ are not inherited vocabulary but local approximations of a northern medieval value like [ʋi], as Pulleyblank was aware.

    Phonetic series mixture like (MC) l- ~ m- prima facie exists, and no positive evidence suggests some early value other than m- for words of the second type. We need to deal with this problem head-on.

    "works phonologically" >> well… (speculative) *ʁ- > (comparative) m- when/under what conditions?
    "came from somewhere at some point" >> well…
    "who can say for sure" >> indeed; this is "OC" in a nutshell. Without evaluating material in terms of what is relatively "sure" — comparative facts are comparative facts, Qieyun facts are Qieyun facts, etc. — we end up in our own personal halls of mirrors.

  26. Chris Button said,

    July 13, 2023 @ 4:59 pm

    Fuzhou Min readings of e.g. characters like "維" or "惟" are not relevant to Chris Button's hypothetical early Chinese development *ʁ- > m-. Such "words" /mi/ are not inherited vocabulary but local approximations of a northern medieval value like [ʋi], as Pulleyblank was aware.

    Unfortunately, you are entirely missing the point. As such, a gentler tone would have been advisable.

    I am talking about phonetic plausibility in terms of how sounds may be associated with one another by speakers in living languages.

    The point is that a fricative with an association with rounding, be that ʁ- or ʋ-, may be associated with m-. I am manifestly not making any claims about inherited vocabulary, northern medieval values or whatever… nor am I reconstructing any values for proto-Min!

    I also in the same post commented on occasional Han-time use of m- to transcribe Sanskrit ʋ-. But, as with the Fuzhou Min comparison, I wasn’t making any statement about how to reconstruct Sanskrit either!

    Why do phonetic (articulatory and acoustic) considerations based on language as it actually works often go by the wayside in historical linguistics? Practitioner background perhaps?

    Phonetic series mixture like (MC) l- ~ m- prima facie exists, and no positive evidence suggests some early value other than m-

    And what to do about the cases where comparative evidence (be that Tai loanword, Tibeto-Burman cognates, internal evidence, etc.) show that a cluster cannot be reconstructed?

    The classic example is 卯, as I pointed out based in Li Fang-kuei’s observations.

    The “proof” you are looking for will come in the “pudding” of a dictionary I am working on rather than cherry picked examples and discussions on theory. Force of numbers, so to speak.

    In the meantime, to take another couple of classic examples in addition to 卯 (i.e., I’m not plucking these out of the air), solve 貉 EMC ɣak and 貉 EMC maɨkʲ (loaned for 禡 maɨʰ), or take a stab at the relationship between 朝 and 廟 for me…

  27. Chris Button said,

    July 13, 2023 @ 5:28 pm

    [ʁʷ] or [ʋ], tomayto or tomahto

  28. Doctor Science said,

    July 13, 2023 @ 9:05 pm

    @Jonathan Smith:

    Silk domestication probably started c. 7500 years ago [5500 BCE] and was completed c. 4000 y.a. [2000 BCE]. I don't know what evidence there is about languages at that time, how broad an area proto-Old Chinese or proto-Sinitic covered, and what the alternatives might be.

  29. Jonathan Smith said,

    July 13, 2023 @ 9:50 pm

    Thanks @Doctor Science; as it happens I was just (trying to) process this one on peach (also apparently a Yangtze domesticate); looks like these articles appeared within a couple weeks of each other… 'silk(worm)' is definitely a promising research project for someone, though linguistic evidence that speaks directly to the relevant period will be hard to come by :D

    @Chris Button. I really have to try harder to accept that your approach such as it is simply reflects the dubious nature of "OC scholarship" period. BUT for the moment maintaining the pretense that historical linguistic methodological norms apply to "OC" as well: whatever other questions may or may not be pertinent, available evidence points plainly to *m- in early Chinese 'wheat' as well as in perhaps-phonologically-comparable '4th Earthly Branch' (here including all Tai material, as I guess you know but have avoided stating.) What the exact situation was at the (early 2nd millennium BCE?) time when the associated written characters were coined is of course harder to say but probably needs to riff off of the comparative facts. Yes I have misunderstood your discussion of Fuzhounese; I thought you were pointing to support there for your proposed sporadic pre-proto-Chinese phonological development ʁ- > m-.

  30. Chris Button said,

    July 14, 2023 @ 10:09 am

    @ Jonathan Smith

    '4th Earthly Branch' (here including all Tai material, as I guess you know but have avoided stating.)

    On the contrary, I have explicitly cited Li Fang-kuei’s comment about the problems with a cluster onset for 卯, which he notes to be “extremely doubtful” on the basis of Tai evidence,

    So, what’s your suggestion?

  31. Chris Button said,

    July 14, 2023 @ 9:56 pm

    Pulleyblank "Old Chinese *-s" (1973) notes how OC -s occasionally hardened dialectally to Middle Chinese -t (the two readings of 四 as EMC siʰ ~ sit being the classic example).

    Looking at the examples more closely, it seems like this shift was conditioned by a palatal component. So, for example, we have 鼻 and 比 (in one of its many readings) giving EMC bjiʰ ~ bjit. The idea that the final -s in OC *-ǝˑjs (perhaps earlier -jǝˑs if the -s was a coda rather than a suffix) could sometimes harden to -t makes good sense. Its also worth noting that the reflex of *-ǝˑjs (or indeed -ǝˑcs) then matches that of -ǝˑc in such instances.

    The problem with 糲 is that we have *-as giving lajʰ ~ lat. However, it turns out that the -s in Proto-Austronesian *bəʀas represents a palatal fricative -ç in the reconstruction proposed by Blust. That palatal component could then account for the -t reflex in Old Chinese as a loanword.

    @ Victor Mair & Martin Schwartz

    Prior to Sanskrit, it is speculated to be either a borrowing from a Dravidian language (compare Proto-Dravidian *wariñci (“rice”)

    I suppose Austronesian *bəʀaç with a palatal coda now starts to look somewhat similar … possibly?

    … the reconstruction I saw in Boodberg which looked like it was compatible with PIIr. *wrinźh

    Boodberg (p.171) wants to compare 米 (EMC mɛj' OC mǝjɁ) rather than 糲/粺/稗 with Tibetan bras and claims an original OC -s coda in 米. The phonology is highly problematic, so I have major doubts. Having said that, since it's fun to speculate and the Tibetan form seems to go back to Austronesian bəʀaç, the following could be noted: ᵐb- is a possible source of EMC m- (although I haven't uncovered any evidence in the phonetic series for it here), -jɁ could be deemed acoustically somewhat close to -ç, and perhaps the uvular trill -ʀ- was just deemed too far removed from OC medial -r- and ignored. It seems too much of a stretch, but who knows?

  32. Jonathan Smith said,

    July 16, 2023 @ 11:52 am

    Re: 'wheat', '4th Earthly Branch'…

    To rehash, all comparative and "Sino-Xenic" evidence points to m- here, so the idea that such words actually begin with ʁ in early Chinese will require extraordinary evidence.

    By contrast, m- presents no problems for the common (Baxter 1992, Schuessler 2007, Baxter/Sagart 2014…) assumption that these words had some complex onset configuration like *mr in "OC" (not "my idea"). Please note cluster > simplex m- in Chinese whence to Tai is an obvious possibility noted by Li Fang-Kuei in that 1945 paper.

    And "1945" — serious study of the Tai evidence will require reference to newer work in comparative Tai (e.g. Pittayaporn 2009) and probably original reconstructions of the Earthly Branches within Tai to the extent this is possible. This was on an older to-do list of mine…

    If anyone is actually interested in MY ideas about things like MC m~l xiesheng mixture, they can look at the CLAO paper on the topic…

  33. Chris Button said,

    July 16, 2023 @ 12:18 pm

    Please note cluster > simplex m- in Chinese whence to Tai is an obvious possibility noted by Li Fang-Kuei in that 1945 paper.

    Let's cite Li in full so as not to misrepresent anything:

    "Anc. Ch. m- (<Arc. ml-) is here represented by m- in all three languages. As Ahom regularly preserves an initial cluster such as ml-, and we also expect a trace of it in Dioi (cf. No.2, where the -l- is vocalized into -i-), the archaic form ml- is extremely doubtful, unless we can show that ml-, different from pl-, had been simplified into m- before it was loaned into the Tai languages. Furthermore the tone in both languages where tones are known to us indicates a 'voiceless initial.'*

    * When we say that the tone indicates a voiceless initial, it simply means that it is a tone ordinarily developed from a voiceless initial. Under special conditions a voiceless initial may give a tone similar to that of a voiced initial and vice versa, cf. Li, op. cit., §9-14."

    serious study of the Tai evidence will require reference to newer work in comparative Tai (e.g. Pittayaporn 2009)

    By contrast, m- presents no problems for the common (Baxter 1992, Schuessler 2007, Baxter/Sagart 2014…) assumption that these words had some complex onset configuration like *mr in "OC" (not "my idea").

    Please could you explain how either of these points affect the evidence presented by Li Fang-kuei?

    If anyone is actually interested in MY ideas …

    I fear we might be losing some objectivity here.

  34. Chris Button said,

    July 16, 2023 @ 2:11 pm

    Please could you explain how either of these points affect the evidence presented by Li Fang-kuei?

    I think I might need to be more specific to avoid any further misunderstandings:

    Of course *mr- is not your idea. Li's point in 1945 (!), where he has ml-, is that it doesn't seem to hold here. Why do you think Pulleyblank dedicated so much attention to this particular case?

    Also, I don't believe Pittayaporn discusses Ahom at all and only mentions Dioi in passing. Those are the two languages Li is referring to. Perhaps I'm missing something?

  35. Jonathan Smith said,

    July 16, 2023 @ 7:00 pm

    Hi Chris,

    I acknowledge your efforts to overthrow the would-be OC orthodoxy re: "mr-" in early 'wheat', etc. — however, you are struggling to come to terms with the fact that comparative + Sinoxenic including Tai evidence points across the board to m- in these words; i.e., no support is found for your own novel idea of an onset ʁ-. Sorry about that.

    Whether '4th Earthly Branch' within Tai can as of 2023 be reconstructed to *m-, *ml, or some other, and at what stage(s) exactly, is of interest but a secondary concern as neither value contravenes conventional OC *mr-. Perhaps a kind Tai specialist will come by and be candle to our darkness on this point.

    To fill LL with posts concerning your own shall-we-say half-baked ideas about early Chinese, formulated as statements of fact to boot ("OC uvulars tended to condition rounding"; "麥 məɨkʲ ← ʁək 'wheat'," etc., etc., ad literal nauseam), only to complain of "loss of objectivity" when others make so bold as to link in response to their own published work on the question at hand, is FYI a comically bad look — but what's new :D

    Until the next bad idea,

    JMS

  36. Chris Button said,

    July 16, 2023 @ 8:19 pm

    uvulars tended to condition rounding

    Maybe take that up with Pan Wuyun or Jin Lixin then, who seem to have followed Pulleyblank's lead there.

    And we actually just discussed on the wheat thread how the "rounding laryngeal" in PIE was quite possibly ʁ-:

    https://languagelog.ldc.upenn.edu/nll/?p=59111#comment-1606924

    It's phonetics and phonology, not algebra.

  37. David Marjanović said,

    July 17, 2023 @ 3:18 pm

    …but quite possibly PIE *o wasn't rounded yet. The latest review is here.

    (Part 1 of a series that has not been continued yet, unfortunately. The author is currently working on his thesis, which is on Proto-Uralic, not PIE.)

  38. Chris Button said,

    July 17, 2023 @ 10:39 pm

    @ David Marjanović

    Out of curiosity, are you personally familiar with the rounding effects supposed to be associated with uvular fricatives in Germanic dialects?

    (I vaguely recall you mentioning your linguistic background elsewhere once on LLog).

  39. Chris Button said,

    July 19, 2023 @ 8:52 pm

    It turns out that Ferlus "Austroasiatic vocabulary for rice" (2010) has proposed that Austronesian *bəʀaç is related to Austroasiatic *C.rac "reap" (with C as a pre-syllabic consonsant). Incidentally, Shorto (1972) "MK contact words in ST" correctly associates Mon-Khmer *rac with Burmese rit "reap" (whose phonological shape betrays its loanword status in Burmese), while also noting a form *rə(ə)c.

    The semantic relationship between "reap" versus "rice" is comparable to say the relationship between Greek karpos "fruit, grain" and keiro "shear".

    In addition to 糲, it turns out that Ferlus also remarks on the similarity between his C.rac and the Dravidian forms that I noted in a comment earlier on from the o.p., which ultimately seem possibly to include the word "rice" itself. Ferlus even speculates that the word "rye" (as I originally talked about here https://languagelog.ldc.upenn.edu/nll/?p=59111) might form part of the picture .

    All this circularity (and the irregularity that comes with loanwords, where one must rely on phonetic plausibility rather than the comparative method for native vocabulary) makes me wonder if 麥 *ʁək “wheat” (and its association with 來 rəɣː “come”) might be similar to 犬 and 狗 for "dog"? Pulleyblank suggested 犬 and 狗 were related phonologically, and Prof. Mair then suggested they may both be loans from Indo-European but at different time depths.

    Could 麥 ~ 來 *ʁək ~ rəɣː and 糲 *raˑs ~ *ras ultimately go back to the same MK *rac ~ *rə(ə)c or AN (bə)ʀaç loanword origin as well, just at different time depths or via different routes?

  40. David Marjanović said,

    July 21, 2023 @ 6:25 pm

    Out of curiosity, are you personally familiar with the rounding effects supposed to be associated with uvular fricatives in Germanic dialects?

    I've encountered the claim, but I suspect this is fake: [ʁ] (invariably for /r/) is a very recent phenomenon in all of them except a village here and there – too recent to have backed (and concomitantly rounded) unstressed -er in all of Swabian or something.

    [χ] is old in Switzerland and northern Germany and doesn't cause any rounding.

    Ferlus "Austroasiatic vocabulary for rice" (2010) has proposed that Austronesian *bəʀaç is related to Austroasiatic *C.rac "reap" (with C as a pre-syllabic consonsant)

    Looks good to me.

  41. David Marjanović said,

    July 23, 2023 @ 2:21 pm

    The strongest argument against *h₃ causing specifically and directly vowel rounding is that nothing else did: the queen's wedding isn't the quoon's wadding, as someone memorably put it on a Wikipedia talk page.

RSS feed for comments on this post